PerceptiLabs is a visual modeling tool for machine learning with a rich user interface to edit, manage, and monitor your machine learning models while designing and training them.
Note: After logging in to PerceptiLabs you will be redirected to the Model Hub where you can create and manage your models.
Tip: Also be sure to check out our UI Overview video on the video tutorials page.
Key Elements of the Perceptilabs UI
The following are the main elements of the Perceptilabs UI:
1.
Model Hub (Overview): lists existing models that have been created in PerceptiLabs and allows you to view and manage those models.
2.
Modeling Tool (Model): the main workspace area where you build your models.
3.
Training : displays the training statistics for the currently-running, or recently-trained model.
4.
Evaluate: displays statistics about your trained model after it has been tested against test data. You can also create new tests here.
5.
Deploy: provides options to deploy your trained model.
6.
Components: provides Components which you can drag and drop onto your model.
7.
Save: saves the changes that you've made to the current model.
8.
Color Scheme and Quick Links: provides the following links:
Color Scheme: allows you to toggle between light and dark color schemes
Chat: opens a chat window so you can ask us questions.
Report a Bug: opens the forums where you can report bugs or find solutions.
Help: provides links to documentation and other resources.
9.
Data Settings: displays the data settings that are used by the model. These were initially set when you first created the model using the Data Wizard.
10.
Login/Logout: allows you to sign in and out of PerceptiLabs.
11.
Run: starts training the current model.
12.
Weights: load and use your most recent checkpoint.
13.
Preview: toggles weather PerceptiLabs displays previews in the model's components.
14.
Settings: toggles display of the settings and components panes down the right.
15.
Open Code: opens the Code Editor for the currently selected component.